Avoid These Common Mistakes When Building Personality into Chat

By Jasmine Mayo, Head of Communications at GameOn Technology

For the last few years, companies have been getting serious about designing unique personalities for their brand-driven chat experiences. For instance, as I’m scrolling through my LinkedIn newsfeed, I see more and more job advertisements popping up for chatbot persona developers, interaction engineers, or conversational designers.

As great as these investments may be at improving customer experience and enhancing profitability, brands still face an uphill struggle to get their chat personalities right. There are many things to consider when building these experiences — How will my customers react and engage? What audience segments use this? Does this experience tell the right story?

These are some of the most important questions brands seeking consumer affinity should be asking themselves. However, months can be lost during the ideation stage, and outcomes of these processes are often suboptimal or overcomplicated.

 

Time and again we see the same mistakes being built into chat functionalities and personalities. These are easy blunders to avoid with a few considerations in mind.

Mistake No.1: Utilizing One Personality Throughout the User Experience

Many brands view chat experiences through the paradigm of customer service. It’s there if you need it when a customer has a problem or query. While this is definitely a major use case and can reduce administration costs as well as improve customer experience KPIs, chat applications can play other roles in the marketing and sales funnel forming a valuable part of a business’s customer engagement.

This is where personality comes into play. Chat experiences that have unique personalities are increasingly becoming a destination in their own right that users are seeking out. This can drive traffic to a brand’s website or social profiles, increase engagements, and convert users into leads and sales.

Why is this important to consider when designing a chat experience with personality? Well, think about the ideal personality traits in a customer service rep versus a sales rep. A great service rep should be patient and empathetic, whereas a high-performing sales rep would be high-energy and assertive. Alternatively, if a user is experiencing a company’s chat application because they are enthused or even angry, they are going to want a different tone from the brand they are engaging with. Therefore, you will need to develop multiple personality traits to better fulfill each of these segmented tasks.

Here’s a case study. We at GameOn developed a chat experience for English Premier League (EPL) soccer team Arsenal. The personality “Robot Pires” was built around a famed former player named Robert Pires. Not only does the chat experience provide fans with the latest results, fixtures, and videos, it also facilitates ticket and merchandise sales. The tone and imagery used by the chat experience differ depending on each of these use case segments.

Mistake No.2: Not Anticipating the Changing Moods of the Users

Users will often be in different states of mind and specific moods when they engage with chatbots. A lot of times this will be impossible to predict. In these situations, well-designed chat experiences will respond appropriately based on the request being made and the language being used by the user.

However, there are a lot of use cases where it’s possible to predict the general mood of the user. The most obvious example is sports fans. If a football team has just lost to a big rival, the mood among fans will be very different compared to if they’ve just qualified for the Super Bowl.

It would be out of character for a sport’s team’s chatbot not to reflect that mood, as they are just as invested in the team as any fan is. This also applies to other use cases as well, not just sport’s teams. For example, if a retailer is having supply chain issues that are delaying deliveries to customers, you can expect a lot of customers will start venting their frustration to the chat application.

For example, when the Vikings win a game, their chat application responds with ‘skol’, a term both beloved and familiar to their fans. On the other hand, if the Vikings did not have optimal performance, the chat experience’s personality reflects this in its tone as well. In this case, when fans chat with the Vikings, they are provided with opportunities to look forward and ask things like: “when do the Vikings play this team next” or “what are the latest videos or news about where the team is going from here.” When this happens, it is important that the applications’ language is optimistic, reflecting the current mood of the fan.

Mistake No.3: Misplaced or Confusing Humour 

Nobody wants to converse with a chatbot that sounds like a robot. Humour is often seen as a solution to this problem and an easy way to make the chat experience more human. The problem with humor though is that it’s subjective and its suitability is dependent on context.

The last thing a frustrated customer looking for a resolution to their problem wants is a badly timed gag. Going back to the example of a human CS rep, they will judge the mood of a customer and the context of the situation before attempting humor, and if they’re dealing with an unhappy customer, then they would most likely avoid it altogether.

Chat personalities should be designed to avoid the overuse of humor during instances where users could be frustrated, such as if they’re inquiring about the status of an order or wishing to cancel a subscription.

And while your chat personality will be unique, and therefore so will its imagined sense of humor, trying to be too subtle or sarcastic can backfire. This is especially the case with chat experiences that have a diverse user base. In these instances, “dad jokes” and play-on-words are good ways to keep the humor broad and unambiguous. Even a bot built for the wildly popular comedy group, Just for Laughs, has to be able to read the room and respond without jokes when the situation warrants.

Mistake No.4: Making the Personality Too Overzealous    

While AI-powered chat experiences have advanced enormously over the last few years, they still cannot resolve every single query. Therefore, users need to enter a chat experience with the right expectations.

This starts by making users fully aware they are talking to AI and not a person. Research shows that the more human a chat experience feels, the more complex the requests and language from users will be.

As users progress through the chat, the language used needs to be helpful and provide appropriate signposting, but not so overzealous so as to create an expectation that it will be able to resolve everything.

When your chat experience can’t resolve a query, you can use this as a moment of brand connection with your customers, such as a GIF or image. Incorporate more than one response to a confusing question, ensuring the personality remains on brand. And above all else, avoid creating journey loops when a query can’t be resolved. Redirecting users back to a happy path that may provide a solution or even point to a customer service representative to follow up live or via email.

Final Thoughts

There are no other tools out there that can scale one-on-one customer engagements in the way chatbots now can. But simply having a functional chatbot on a website is no longer enough. Customers want to engage with chat experiences that are an extension of the brands that they serve. Unique chat personalities deliver this. However – the perfect personality can’t be built overnight, it takes careful planning, testing, then constant iterations once live.

Written by Jasmine Mayo, Head of Communications at GameOn Technology, a team of industry-leading designers, developers, and strategists whose platform and expertise power award-winning conversational AI applications for the world’s most innovative brands.

SIGN UP TO SKILL UP
The All Access membership allows you to discover 500+ hours of best in class thought leadership